Releases
0.2.0
v0.2.0 - Scoring System Refinement
Major Changes
Removed unused neural network components for better performance and transparency
Adjusted difficulty scoring thresholds for more accurate word classification
Completely revamped documentation with detailed scoring explanations
Scoring System Updates
Changed simple/complex word threshold from 0.3 to 0.4
Adjusted sigmoid curve parameters for smoother scaling:
Simple words (< 0.4): Gentler curve with midpoint at 0.2
Complex words (≥ 0.4): Steeper curve with midpoint at 0.6
Rebalanced component weights:
Phoneme complexity: 40% (increased)
Maximum phoneme difficulty: 20%
Syllable structure: 20%
Phonotactic constraints: 10%
Prosodic patterns: 5%
Complex phoneme ratio: 5%
Performance Improvements
Removed BERT model dependencies
Eliminated unnecessary model loading and inference
Reduced memory footprint
Documentation
Added detailed explanations for each scoring component
Included language-specific examples with scores
Improved API documentation with real-world usage examples
You can’t perform that action at this time.